BLAST PDB web services being removed November 2020

[Tom Goddard]

The PDB announced that legacy PDB web services will be discontinued November 2020, including ones used by the ChimeraX Blast PDB tool.

Searching for /pdb/rest in ChimeraX code gives

bundles/blastprotein/src/pdbinfo.py320: url = '​https://www.rcsb.org/pdb/rest/%s?structureId=%s' % (query, ids)

[Tom Goddard]

Announcement from PDB:

---

Subject: Deprecation of RCSB PDB legacy APIs
Date: Fri, 8 May 2020 12:17:21 -0700
From: Jose Duarte <jose.duarte@…>
To: Roland Dunbrack <roland.dunbrack@…>, Andrej Sali <sali@…>, Torsten Schwede <torsten.schwede@…>, tef@…, PDAdams@…
CC: Stephen Burley <stephen.burley@…>, Christine Zardecki <christine.zardecki@…>, Rachel Green <rachel.green@…>

Dear advisors, dear Andrej

This week we have announced that our legacy APIs are going to be discontinued in November 2020, see:

​http://www.rcsb.org/news?year=2020&article=5eb18ccfd62245129947212a&feature=true

The APIs have been around for a long time and we understand that there are many programmatic users that will need to adapt.

We are trying to send notifications as wide as possible and we are planning to send messages to CCP4 and to other lists. If you can think of anyone that we should notify specifically, we would appreciate your feedback. Of course, feel free to also spread the news among your teams and colleagues.

Many thanks and best regards

Jose

---
Jose Duarte
RCSB Protein Data Bank
SDSC, UC San Diego

[pett]

Latest announcement:

We are directly reaching out to previous RCSB PDB API users and PDB-related software providers. Apologies in advance for cross posting.

The RCSB PDB legacy APIs are going to be discontinued in November 2020, see the news article:

​https://www.rcsb.org/news?year=2020&article=5eb18ccfd62245129947212a&feature=true

All programmatic users are strongly encouraged to migrate to the new APIs before November 2020.

New json-based APIs are now available that provide more comprehensive search and fetch capabilities, see ​https://www.rcsb.org/pages/webservices. Highlights include :
A GraphQL interface for the Data API
Full boolean support for Search API
Integrated text + scientific searches (structure search, sequence search) in the Search API
As usual, for any questions you can use the "Contact Us" button at the rcsb.org website.

[pett]

Updated Blast considerably slower because the new REST services don't support batching. Need to rewrite to use RCSB GraphQL API, as per this mail:

Have updated to new RCSB REST APIs. Sometime soonish would be nice to rewrite the code to use their GraphQL API where you can batch the fetches, as per this mail:

We recommend GraphQL API where it’s easier to get all data for a PDB ID e.g.

[​http://data.rcsb.org/graphql/index.html?query={
entries(entry_ids%3A 2FBW) {
rcsb_id
polymer_entities {
rcsb_id
rcsb_polymer_entity_container_identifiers {
asym_ids
auth_asym_ids
}
}
nonpolymer_entities {
rcsb_nonpolymer_entity_container_identifiers {
asym_ids
auth_asym_ids
}
}
}
}

​http://data.rcsb.org/graphql/index.html?query=%7B%0A%20%20entries(entry_ids%3A%20%5B%222FBW%22%5D)%20%7B%0A%20%20%20%20rcsb_id%0A%20%20%20%20polymer_entities%20%7B%0A%20%20%20%20%20%20rcsb_id%0A%20%20%20%20%20%20rcsb_polymer_entity_container_identifiers%20%7B%0A%20%20%20%20%20%20%20%20asym_ids%0A%20%20%20%20%20%20%20%20auth_asym_ids%0A%20%20%20%20%20%20%7D%0A%20%20%20%20%7D%0A%20%20%20%20nonpolymer_entities%20%7B%0A%20%20%20%20%20%20rcsb_nonpolymer_entity_container_identifiers%20%7B%0A%20%20%20%20%20%20%20%20asym_ids%0A%20%20%20%20%20%20%20%20auth_asym_ids%0A%20%20%20%20%20%20%7D%0A%20%20%20%20%7D%0A%20%20%7D%0A%7D%0A]

[pett]

Updated Blast Protein to use GraphQL, so all information is retrieved in a single query (instead of hundreds).